The only issue we reach on this appeal is whether the evidence presented by the defendants in this patent infringement action is sufficient to support the district court's finding that their variable resistor was "on sale" in the United States, within the meaning of 35 U.S.C. § 102(b), prior to March 16, 1969. That finding was made after a retrial of the on-sale issue pursuant to our remand in CTS Corp. v. Piher International Corp., 527 F.2d 95 (7th Cir. 1975), Cert. denied, 424 U.S. 978, 96 S. Ct. 1485, 47 L. Ed. 2d 748 (1976). We hold that the evidence was insufficient and therefore reverse.

The facts leading to this litigation and the technology involved are discussed in our prior opinion and will not be repeated here, except as necessary to understand the on-sale issue. Variable resistors, or potentiometers, are "used to adjust the electrical resistance of an electronic circuit to a desired level. The volume and tone controls on a television or radio are examples of variable resistors." CTS v. Piher, supra, 527 F.2d at 97. Variable resistors are of two types: (1) customer adjustable (examples of which are the volume or tone controls just described), and (2) factory set. It is the latter type, sometimes referred to as a "trimmer," that is involved in this case.

In 1972, CTS Corporation filed suit against Piher Sociedad Anonima and its American subsidiary, Piher International Corporation,1 alleging that their PT-15 resistor infringed several CTS patents, two of which were Patent No. 3,518,604 on "Electrical Component" issued June 30, 1970, and Patent No. 3,670,285 on "Variable Resistance Control with End Collector" issued June 13, 1972. See CTS v. Piher, supra, 527 F.2d at 97-98 & nn. 1-3 & 11. Although CTS withdrew its claim that the PT-15 infringed the '285 patent, Piher counterclaimed, charging that both patents were invalid. Id. at 98. In the first trial of this case, the district court upheld the validity of both patents and found that the PT-15 infringed the '604 patent. On appeal, we affirmed the judgment as it related to the '604 patent. Id. at 101. But because we found that the district court had erroneously excluded evidence supporting Piher's contention that the '285 patent was invalid under § 102(b), we remanded the case "for a new trial of all issues raised by Piher's on sale challenge to the validity of the '285 patent." Id. at 105.

At the first trial, Piher introduced evidence and made offers of proof tending to show that in the fall of 1968 Piher had shown prototypes of the PT-15 resistor to the manager of International Materials for Motorola, Inc. Id. at 102. Further, there appeared to be evidence that Motorola ordered a trial run of these resistors on January 22, 1969, that Piher shipped some resistors on March 3, 1969, and that Motorola received some resistors in April, 1970. After receiving the resistors in April, Motorola tested them and found them operable. There was evidence, however, showing that Piher's PT-15 had been modified from time to time, and "(t)hus, the character of the devices shipped by Piher on March 3, 1969, was of critical importance." Id. at 103. This court expressly declined to make the requisite findings of fact because "the question of credibility and the interpretation of the exhibits are matters that must be decided in the first instance by a trial judge." Id. at 105.

On remand, the district judge, making his findings orally, and apparently extemporaneously, immediately following the conclusion of the trial, found that Piher had carried its burden of proving that a PT-15 which disclosed the essential elements of the '285 patent, Id. at 103, was "on sale in this country more than one year prior to the application for patent in the United States," 35 U.S.C. § 102(b), and therefore the '285 patent was invalid. CTS's assignor applied for what became the '285 patent on March 16, 1970. CTS v. Piher, supra, 527 F.2d at 97 n. 2. Thus the critical date for establishing the § 102(b) claim of invalidity is March 16, 1969.

The governing rule laid down in CTS v. Piher is that it was unnecessary for Piher to prove a completed sale of the PT-15 in the United States prior to March 16, 1969, in order to establish the on-sale defense under § 102(b), and that if before that date Piher was

offering to prospective purchasers in the United States a product which (a) embodied the '285 invention and (b) was complete in the sense that it represented a reduction of the invention to practice, the invention was on sale within the meaning of the statute and the '285 patent is invalid.

527 F.2d at 103. CTS makes certain arguments concerning the proper interpretation and even the correctness of our earlier decision which we need not address. We assume for purposes of this case that it would be enough that the invention was reduced to practice in Spain before the critical date even if the initial offer pursuant to which the prospective customer placed an order was made before the reduction to practice. Our decision is based upon our conclusion that the evidence was insufficient to support a finding that Piher established reduction to practice in Spain before March 16, 1969.

We made it clear in our prior opinion that Piher had the burden of proving reduction to practice before the critical date. Id. at 103. We noted a fact that made this burden more onerous, Viz., "there is evidence in the record that the trimmer has been modified from time to time." Id. We also indicated that the burden must be met by clear and convincing evidence. Id. at 105; See also Red Cross Manufacturing Corporation v. Toro Sales Company, 525 F.2d 1135, 1139 (7th Cir. 1975).

Reduction to practice may be established without proof that the device would have been commercially successful, if marketed. Dart Industries v. E. I. DuPont De Nemours, Inc., 489 F.2d 1359, 1365 & n. 11 (7th Cir. 1973). Nevertheless, the device must be shown to have been tested and found operable.

The district judge made two findings on dates of reduction to practice: He found that the PT-15 had been reduced to practice by Piher in Spain in November 1968; and he found separately that completed devices were in existence in Piher's possession in Spain on about March 3, 1969. If either finding were sustainable, the requirement of reduction to practice would be satisfied. We therefore examine both findings.

Before addressing the evidence relevant to each finding, we note a circumstance that weighs heavily against Piher and diminishes the force of the evidence it introduced. We refer to the unexplained absence of the kind of evidence we would normally expect to see in a case of this kind, Viz., Piher's records of the development,2 testing, and production of the device Piher alleged it had reduced to practice before the critical date. The district court refused to draw any inference adverse to Piher from this circumstance, apparently because of uncertainty as to record-making and record-keeping practices in Spain.3 If Spanish business practices were the reason for Piher's failure to produce the records, it was for Piher to so demonstrate, not for the court to assume the fact without proof. Piher's failure either to produce the records or to offer an explanation of why they could not be produced gave rise to an inference that they were unfavorable to Piher. See 2 J. Wigmore, Evidence §§ 285, 291 (3d ed. 1961 & Supp. 1977). This inference is especially telling in view of Piher's burden of proving reduction to practice by clear and convincing evidence.

If the alleged reduction to practice in Spain occurred before the critical date, it was achieved by Piher employees. The absence of documentation was compounded by Piher's failure to offer testimony by any person who had participated in, or otherwise had direct knowledge of, the alleged reduction to practice in Spain before the critical date or to explain why testimony of this nature was unavailable. The silence from Spain leads to an inference unfavorable to Piher. 2 J. Wigmore, Supra, § 286.

We turn now to the evidence Piher did offer.

In support of its finding that the PT-15 was reduced to practice in November, 1968, the district court relied almost exclusively on the testimony of Joel Adams, who was Motorola's European buying manager between 1966 and 1970. Adams testified that in November of 1968 he visited Piher's manufacturing plant near Barcelona, Spain, where he was shown samples of the PT-15 that "looked" complete. Adams also stated, however, that he did not "test" the samples or even examine their internal construction. And further, in response to a question from the court, Adams frankly stated that he "couldn't attest that they really worked electrically because I didn't test them."

The evidence was clear that it was not until April, 1969, that Motorola received completed samples of the device the court found was complete in November, 1968, notwithstanding Motorola's repeated requests for such samples beginning in October, 1968.4 The court found this evidence of receipt in April insufficient to support a conclusion that the samples had been shipped before March 16, 1969, but failed to make any finding as to when they were shipped.

Piher introduced in evidence as Defendant's Exhibit 48 one of the PT-15 variable resistors that the court found had been received by Motorola by April 7, 1969, and tested and found operable by Motorola in April, May, and June of 1969. The court found that this group of resistors, including Defendant's Exhibit 48, were "typical of those devices" shown to Adams in November of 1968. Yet the court acknowledged the force of CTS's argument that if the PT-15 was complete in November, 1968, it was unlikely that Piher would have failed to ship them until after March 16, 1969:

(I)f the devices were available as early as November of 1968, why were they not given to Motorola at that time? Why was there so much difficulty getting samples? And why did they not arrive until April of 1969?

I will tell you frankly that I think that is the strongest part of the plaintiff's case, and I have thought long about that.

Adams testified that the resistors shown in the fall were not of the proper "value," and Piher argued that this explained why no shipments were made until the next spring. The court, however, characterized as "telling" CTS's argument "that these devices were manufactured in the first instance specifically for Motorola and there is no reason to think they would not have been of the proper value," stated that it had "difficulty" with the question, but nevertheless proceeded to find without further explanation that the devices were completed in November but were not of the proper value.

Motorola's October, 1968, request for working samples indicated an interest in resistors in a range of specified values. So far as the record shows, Motorola was Piher's only sales prospect at that time. As early as September, 1968, a meeting with Motorola representatives had been scheduled for November. In the absence of any explanation, it seems to us unlikely that Piher, which had planned for the meeting since September and was aware in October of the values Motorola was interested in, would produce complete resistors to be shown to representatives of Motorola in November with values different from any value Motorola had specified.

During the period in question Ricardo Balil, president of Piher's American subsidiary, regularly sent Telexes to Piher in Spain reporting on his selling activities in the United States. One of these reports, dated February 17, 1969 relates to a prospective customer other than Motorola and contains the following statements:

Please send two units of each of the PT-15. It doesn't matter value. They are interested in the different types of assembly or mounting.5

Ten days later, apparently referring to this and other similar requests by Balil, Piher responded with the statement that when the devices were in production, which would be in about two or three weeks, Piher would send some of the devices to Balil. The record does not show when, if ever, this promise was fulfilled. Nor does the record contain any explanation of why, if the PT-15 was complete in November and the only problem was the proper value, as the court found, Piher did not promptly send Balil a few complete samples of any value, in compliance with his request of February 17.

In view of the events just described, it is not surprising that the district court was not persuaded that Balil had brought resistors identical to the ones received by Motorola in April, 1969 from Spain in December of 1968, despite his testimony that he had done so. The court also refused to accept Balil's testimony that he had shown "completed samples" to prospective customers in the United States during January and February, 1969, although if completed samples had been available it is difficult to believe Balil would not have had some to show.

We conclude that Piher fell far short of proving by clear and convincing evidence that its PT-15 was reduced to practice in November, 1968.

We turn finally to the district court's finding that the PT-15s received by Motorola on or about April 7, 1969 were complete on or about March 3 of that year.6 That finding stands in contrast to the court's expressed conclusion that the evidence was insufficient to show the devices were shipped before March 16, 1969. The relevant evidence consists of three documents: (1) a packing slip dated March 3, 1969 listing quantities of four types of PT-15s described by Piher's part number but not serial number or other reference identifying any specific device;7 (2) a memorandum of the same date from the planning department of Piher's Spanish plant to Balil, presumably, but not explicitly, referring to the same group of PT-15s listed in the packing slip and stating that they were "attached";8 and (3) a letter from someone at Piher's Spanish plant to Adams' assistant at Motorola dated March 7, 1969, stating that Motorola's purchase order number 295801,9 which was for a total of 1100 resistors, 110 of each of ten types, had been filled and shipped on March 6, 1969.10 To the extent that the letter represents that the entire order had been filled and shipped, it is incorrect. Adams' testimony that the first PT-15s received by Motorola in April were only a partial shipment and that the balance of the order was filled in the succeeding months, is uncontroverted. On these three documents, which the court considered insufficient to establish that the PT-15s received in April were shipped on or before March 16, 1969, rests the finding that completed PT-15s existed almost two weeks before that date, on or about March 3.

The district court found that it was "more probable than not that the shipment of potentiometers received in the United States and delivered to (Motorola) on or about April 7th was the very shipment referred to in" the packing slip. After expressing considerable doubt on the point, the court stated that it was "inclined to believe that the shipment represented by" the packing slip was what Motorola received by April 7 because there was "no evidence of any other shipment being received at about that time."

It is apparent that the court applied a lower standard of proof than required under § 102(b), but, apart from that, we find it difficult to reconcile the court's inability to find from these documents that the samples referred to in the packing slip were shipped between the 3d and 6th of March, as the memorandum and letter recited, with the finding that the samples were completed on March 3d. We have already noted that Motorola had been pressing for the samples for months, and Piher had been promising that they would soon be available.11 In view of this urgency, it seems to us less likely that if samples had existed they would not have been shipped than that the documents were the products of hope or misunderstanding, or both. We have noted that the letter was demonstrably inaccurate in one respect: the majority of the devices ordered were not shipped until much later. The evidence shows other instances in which Piher's documentation of shipment proved to be inaccurate. In view of the likelihood that the samples would have been sent as soon as they were completed, we cannot understand how the same evidence that failed to establish that the samples were sent before March 16th did establish that they were completed before that date.

The district court clearly did not rely upon the only possible basis other than the three documents for concluding that the PT-15s received by Motorola in April were reduced to practice before March 16, 1969, Viz, the time required in 1969 for a package shipped air parcel post from Spain to reach the United States. The only evidence as to transit time was Balil's estimate of "two to three weeks," which the district court expressly rejected, presumably because it had found Balil unreliable as a witness on other points. The court "guessed" that it would take about seven to ten days. If the court's guess was accurate, the PT-15s received on April 7 might not have been completed and tested until March 29, 1969, after the critical date.12

Giving appropriate weight to the inferences that must be drawn from the unexplained absence of direct evidence of reduction to practice, we cannot say that the record supports a finding, based on "clear and convincing" evidence, that the PT-15s received by Motorola on April 7, 1969, were complete and proved operable by tests conducted in Spain prior to March 16, 1969. The district court's finding on this point was clearly erroneous. Accordingly, the judgment is reversed, and the case is remanded for the entry of an appropriate judgment in favor of plaintiff.

REVERSED AND REMANDED.